Introduction
OpenAI has released GPT Image 1.5 today, a critical update to its image generation pipeline, in what appears to be a direct and necessary response to the recent dominance of Google Nano Banana Pro and the steady rise of Midjourney V7. Following weeks of speculation about an internal "Code Red" declared by Sam Altman due to lost market share (estimates suggest DALL-E 3 usage dropped by up to 80% in some verticals in favor of Flux and Google), this release bets everything on three pillars: speed (4x), granular control, and cost efficiency.
Available immediately for all ChatGPT users and via API, GPT Image 1.5 doesn't just aim to improve aesthetic quality but to solve the number one problem of generative AI in 2025: editing consistency.
Analysis and Technical Details
The core of the update lies in the model's ability to act as an intelligent editor rather than a purely stochastic generator.
Precision and Instruction Following
Unlike previous versions (and many competitors like early Flux iterations), GPT Image 1.5 introduces an architecture capable of isolating semantic elements. If a user asks to change a subject's expression or a scene's lighting, the model rigorously preserves details such as facial features, branding, and original composition. This level of intent preservation is critical for enterprise workflows where brand consistency is non-negotiable.
New 'Images' Workspace
The software update includes a new dedicated interface in the ChatGPT sidebar (mobile and desktop). This space acts as a "creative studio," offering:
- Preset filters to reduce prompt engineering.
- Trending prompts updated regularly.
- Non-destructive editing: ability to add, subtract, or blend elements without regenerating the entire image.
API and Economics
For developers, the biggest news is the price cut. The GPT Image 1.5 API is 20% cheaper than GPT Image 1. This move is clearly defensive, aiming to win back enterprise customers who migrated to open-source self-hosted solutions like Black Forest Labs (Flux) or Google's aggressive Vertex AI pricing.
Market Impact and Competitors
Today's release (December 2025) redraws the competitive map:
- Vs. Google Nano Banana Pro: Google's model has dominated the last quarter thanks to its virality and photorealism. OpenAI responds by focusing on business usability and precise editing, areas where Nano Banana is still perceived as "chaotic" or overly creative.
- Vs. Midjourney V7: While Midjourney V7 remains the undisputed king for pure artistic quality, its workflow (often still tied to Discord or complex web apps) is less integrated than the ChatGPT ecosystem. GPT Image 1.5 doesn't beat V7 on art, but it wins on productivity and textual editing.
- Vs. Flux (Black Forest Labs): Flux has eroded huge market share due to its open-weight nature. OpenAI's 20% API price cut is an attempt to stem this bleeding, making proprietary GPU server management less attractive for mid-sized companies.
Conclusion
GPT Image 1.5 is a substance-over-style update. It abandons the race for mere pixels to focus on real utility: editing an image without breaking it. In a 2025 where image generation is a commodity, the real battle is over control. With this release, OpenAI proves it has listened to post-DALL-E 3 criticisms, finally offering a tool that can integrate into professional workflows without the frustration of randomness.
FAQ
When was GPT Image 1.5 released?
GPT Image 1.5 was released on December 17, 2025, and is available immediately for all ChatGPT users and developers via API.
Is GPT Image 1.5 better than Google Nano Banana Pro?
It depends on the use case. Nano Banana Pro is often considered superior for photorealism and virality, while GPT Image 1.5 excels in editing consistency and following complex instructions.
How much does the GPT Image 1.5 API cost?
The GPT Image 1.5 API costs 20% less than the previous generation (GPT Image 1), making it more competitive for enterprise integrations.
Can I edit images while keeping faces the same?
Yes, a key feature of GPT Image 1.5 is the ability to preserve critical details like faces, logos, and lighting during edits (inpainting).
How do I access the new Images workspace?
The new dedicated "Images" space is accessible directly from the sidebar of the ChatGPT mobile app or on chatgpt.com.